The very sluggish recovery of the economy since the financial crisis — despite zero and near zero interest rates — presents the dominant school of New Keynesian macroeconomists with a conundrum. Many have attempted to resolve the riddle by arguing that such unprecedentedly low interest rates are not the doing of the Fed and therefore do not indicate an expansionary monetary policy. Although not formally a New Keynesian, George Selgin has taken up and vigorously defended this position. According to Selgin, the view that interest rates have been “held down” by “the Fed’s easy money policies” is based on a “myth.” “The unvarnished truth,” according to Selgin, “is that interest rates have been low since the last months of 2008, not because the Fed has deliberately kept them so, but in large part owing to its misguided attempt, back in 2008, to keep them from falling in the first place.” Indeed, in Selgin’s view, the Fed’s monetary policy actually has been “too tight” since 2008.
Let us analyze Selgin's argument, which consists of a number of empirical and theoretical claims. We’ll start with his empirical claims. Selgin contends that the policy of “quantitative easing” (QE) “represented an easing of monetary policy only in a ceteris paribus sense.” That is, QE would have expanded the money supply had it not been neutralized by other Fed policies. These policies include the payment of interest on excess reserves (IOER) and the Treasury’s Supplementary Financing Program (SPF), which either increased the demand by commercial banks for the reserves that the Fed was creating (IOER), or funneled them into a special Treasury account held at the Fed (SPF). Now it is certainly true that in theory these programs could offset or even reverse the expansionary effect of QE on the money supply. But it is easy to determine the actual effect of these programs by simply examining the data on the growth rates of monetary aggregates since 2008. Curiously, rather than following this obvious and simple procedure, Selgin presents a single chart showing the changes in total deposits at Federal Reserve banks held by the Treasury under the Supplementary Financing Account, commenting, “At one point . . . the SPF program alone immobilized almost $559 billion in base money preventing it from serving as a basis for private-sector [i.e., fractional-reserve bank] money creation.” But Selgin’s chart shows that this large neutralization of reserves only occurred for a few months in 2008, and never sidelined more than $200 billion in reserves from 2009 until the early 2011 when the program was terminated. More important, this chart gives us no indication whatsoever of the net effect of the combination of QE and the countervailing programs on monetary growth.
In fact, as we can see from Chart 1, for the nearly six years from mid-2011 to 2017, the year-over-year (YOY) growth rates of M2 and MZM varied between 5% and 10%. Selgin does concede that the IOER policy failed to prevent the effective fed funds rate from declining to the “zero bound,” although he counters that it did succeed in encouraging banks to hoard some of the newly created reserves instead of using them to purchase assets and thereby create new money. But once again the question must be asked: why doesn’t Selgin just directly examine the variations in the growth rates of the money supply since 2008? It is noteworthy that the rates of monetary growth during the later period are comparable to and may slightly exceed the rates during the run-up of the housing bubble from the beginning of 2002 through 2005.
Another one of Selgin’s empirical claims that can easily be tested against the data is that there is no evidence that the Fed has been following an “easy monetary policy,” because monetary ease must lead to “an eventual increase in nominal spending, if not the rate of inflation. Yet, as everyone knows, neither of these things happened.” Selgin then goes on to display charts showing that GDP growth was negative between September 2008 and the same month in 2009 and that the inflation rate fell to either 1.00 percent or into negative territory (if we exclude food and energy) for six months beginning with March 2009. Yet his charts take us only to the end of 2009, which hardly tests Selgin’s claim that the Fed did not pursue a policy of monetary ease because an “an eventual increase” in GDP, i.e., nominal spending, and inflation never occurred. (Emphasis added.)
As Chart 2 shows, almost immediately after the period that Selgin considers, the YOY growth rate of GDP spurted up to nearly 5%. Between 2010 and 2017 it fluctuated in a range between 2.5% and 5.0%. By Selgin’s standards, this is surely evidence of an expansionary monetary policy. Indeed, in his book, Less Than Zero, Selgin (1997, pp. 64-66) calls for stabilizing the growth rate of nominal GDP at 0% per annum, thus allowing the price level to naturally decline in response to increases in labor (or total-factor) productivity induced by technological progress and capital accumulation.
And, indeed, as we see in Chart 3, outside of the period encompassing the end of the Great Recession and its immediate aftermath, the only period for which the CPI was at or slightly below zero occurred in the first nine months of 2015, when oil prices tanked. For most of the rest of the period the inflation rate fluctuated between one 1% and 2%, with two multi-month spikes into the 2%-3% range and one spike into 3%-4% range.
We should also note that positive inflation rates occurred in the face of a sustained fall in velocity of the monetary aggregates M2 and MZM which began in 2006, as shown in chart 4. Had the Fed merely offset this “demand-side shock,” to use New Keynesian terminology, as Selgin urges in Less Than Zero, then the inflation rate should have been negative to reflect growth of labor productivity at an average annual rate of 1.2% (chart 5). Thus the U.S. economy since 2009 has certainly experienced “relative inflation,” which Selgin (1997, p. 55) defines as “output prices rising relative to unit costs.” But Selgin gives no explanation of how such a relative — and for most of the period, absolute — inflation could develop and be sustained for seven years absent expansionary monetary policy by the Fed.
Given these data, we must therefore reject Selgin’s empirically-based conclusion that the Fed was not engaged in expansionary monetary policy after the financial crisis and that its
... unprecedented asset purchases, which might ordinarily have been expected to result in roughly proportional increases in broad money, spending, inflation, and nominal interest rates, affected those variables only modestly, if at all, and did so for the most part by limiting their tendency to decline, rather than by raising them in an absolute sense.
Broad money, nominal spending, and prices did undergo a sustained and progressive rise in absolute terms during a period when velocity was steadily declining and labor productivity was increasing, which according to Selgin himself indicates a monetary easing.
In addition to the empirical flaws in his case, Selgin dismisses the application of the “(relatively) tried and true” analysis of the central bank’s policy of driving down the interest rate below its natural level. This is Wicksell’s analysis of the cumulative process and Selgin seems to be confused about the empirical implications and the conceptual foundations of the theory. Regarding the empirical implications, Selgin quotes Larry White:
If the central bank wants to keep the market rate low in the face of the nominal income effect, it must accelerate the monetary injection. Short-term real rates have been negative , and nominal rates near zero, for eight years now, with little signs of accelerating broad money growth or a rising inflation rate.
Based on this reasoning, White, with Selgin presumably in accord, dismisses “the Wicksellian cumulative-process scenario ... as a viable candidate for explaining why current rates have remained so low since 2008.”
Now, White’s description of the cumulative process does not accord with Wicksell’s. For Wicksell, the continuation of the process does not require “accelerating monetary growth,” which implies “a rising inflation rate.” The claim that Wicksell (2007, pp. 196, 201) makes is much more modest:
T]he rise in prices, whether small or great at first, can never cease so long as the cause which gave rise to it continues to operate; in other words so long as the loan rate remains below the normal rate . . . . A lowering of the loan rate below the natural rate ... in itself tends to bring about a progressive rise in all commodity prices.
Elsewhere Wicksell (p. 148) comments on his model of the cumulative process: “It is possible in this way to picture a steady, and more or less uniform, rise in all wages rents, and prices (as expressed in money).”
Thus, in Wicksell’s analysis, the divergence between the two rates implies only a cumulative rise in the price level and thus in the level of the money supply at a “steady” rate, and not necessarily a continual rise in the rate of inflation and the rate of monetary growth. In his presentation of Wicksell’s model, Carl Uhr (pp. 235-41) demonstrates that the cumulative process can continue indefinitely with a constant 1.00% per year increase in nominal income and in the price level. Thus, White’s “nominal income effect” requires only a level change in the quantity of money and prices, and not a rate change in these variables. This is clear in the “one reservation” that Wicksell expressed about his model, according to Uhr (p. 241): “namely, that the entire sequence was predicated on the assumption that entrepreneurs and others act and react only to prices current in their planning periods.” It is only when inflationary expectations are introduced, according to Wicksell, that “the actual rise will become more and more rapid.” We may conclude, then, that the dynamics of Wicksellian cumulative process are completely consistent with the data presented in the charts above.
This brings us to Selgin’s view that the natural rate is fundamentally unobservable and must be inferred from “a mass of empirical studies.” Thus Selgin cites a graph referred to by Janet Yellen which indicates that the natural rate has been negative since 2008. But this is a case of mistaken identity. For the rate that Selgin identifies is not Wicksell’s natural rate but Keynes’s concept of the “neutral” or “optimum” rate. In fact, Keynes explicitly rejected the Wicksellian natural rate as not “very useful or significant.” Unfortunately, today, the terms “neutral rate” and “natural rate” are used interchangeably to designate the rate that was considered of policy significance by Keynes. When Bernanke, Krugman, Yellen, and other New Keynesians refer to the natural rate, they have in mind the interest rate that is consistent with full employment of resources at some targeted, non-accelerating inflation rate. The goal of the central bank is to discover and establish this fictional rate in financial markets, which will in turn drive investment spending and the real rate of return on investment to levels consistent with stability of the real economy.
This New Keynesian notion of the natural rate contrasts sharply with Wicksell’s conception. According to Wicksell (p. 205), who was a follower of Böhm-Bawerk and an Austrian capital theorist, “the natural rate of interest [is] the real yield of capital in production.” The natural rate is thus an “intertemporal” price, or the ratio of prices between present consumption and future consumption (as embodied in capital goods), and it is wholly and directly determined by capital investment in the real sector of the economy. The loan rate of interest is therefore a reflection of the natural rate. As Wicksell (p. 192) put it: “That loan rate that is a direct expression of the real rate, we call the normal rate.” This “normal” or “natural” loan rate derives from the natural rate of return on investment throughout the economy’s capital structure and moves in near lock-step with it: “The rate of interest at which the demand for loan capital and the supply of savings exactly agree ... more or less corresponds to the expected yield on the newly created capital.” (Most of this paragraph is drawn from an earlier publication of mine.)
There is thus no need to undertake econometric and other empirical studies to determine the natural rate. The natural interest rate is nothing but the basic or long-run rate of return on investment in the real structure of production. This fundamental or, what Mises called, “originary” interest rate governs the rate of interest on financial markets, not the other way around, as Keynes and his modern disciples would have it. For Wicksell and the Austrians, it is the real economy dog that wags the financial sector tail. Consequently, any and all attempts by central banks to lower the interest rate via monetary policy inevitably create a divergence between the actual and natural interest rates and initiate Wicksell’s inflationary cumulative process. A complete cessation of Fed open market operations would soon enough allow the underlying interest rate on all financial instruments to return to its natural level in line with the basic rate of return on real investment as dictated by people’s voluntary consumption/saving preferences.
But what of Selgin’s and the New Keynesians’ assertion that the notional natural rate itself has plunged through the zero bound and, therefore, the inflationary Wicksellian cumulative process does not apply, because the Fed does not yet have the tools to push the nominal rate far enough below zero. First, this assertion is absurd on its face, because it is tantamount to the claim that capitalists are investing in real capital goods at a negative rate of return, despite the existence of the universal law of time preference.
Second, we do not need “a mass of empirical studies” to confirm that the natural rate has not plunged into negative territory and may have even risen above its pre-crisis level. Consider the chart below, which appears in a publication by the U.S. Bureau of Economic Analysis (BEA) and is constructed from data in the U.S. national income and product accounts (NIPAs). The top panel plots annual average before tax and after tax rates of return for U.S. nonfinancial corporations for the period 1960-2015. These rates are calculated as the ratio of the net surplus of the corporation to its net stock of produced assets (i.e., capital assets plus inventories valued at current cost). The numerator of the ratio is net operating surplus which is the sum of corporate profits and a few minor items. Corporate profits are a composite of what the economist would call pure or entrepreneurial profit and the return to capital investment (the postponement of consumption). Most of “corporate profits” consist of the normal or natural return to capital investment since pure profits net to zero in a “stationary” or no-growth economy and are slightly positive in the slowly progressing U.S. economy, where saving, investment, and real output per capita is growing slowly. Note that the after tax average rate of return hit a decadal high of 7.6% in 2006 and then fell for the rest of the decade to a low of 6.2% in 2009. It then rose sharply in 2010 to 7.9% and has remained at 8.0% or above through 2015. These variations certainly do not bespeak a collapse of the natural rate after the financial crisis.
The same story is told by a data series calculated by the BEA from industry economic accounts (IEAs), which consists of average annual rates of return for the 71 industries that account for all U.S. economic activity. These return ratios are calculated as net operating surplus divided by the net stock of produced assets for each industry sector. Since the rates of return are calculated for entire industries, the numerator includes both corporate profits and the income of sole proprietorships and partnerships, so there is a significant wage component that inflates rates of return. But it is the variations in the rates of return that are significant. For this series, which is not charted, the decadal rate of return peaks in 2005 at 14.1% and then plunges to 11.7% by 2009, after which it rises rapidly to 13.3% in 2010 and fluctuates between 13.0 and 13.6% through 2015.
To summarize: George Selgin makes three strong, empirically testable claims. First, under current conditions the Fed is incapable of controlling interest rates. Second, the Fed itself is responsible for its own impotence because its monetary policy has been “too tight” since 2008. Third, zero and near-zero interest rates are not indicative of expansionary monetary policy but of a Fed-induced collapse of the natural interest rate to less than zero by tight-money policy. Based on the data adduced above in conjunction with a proper understanding of Wicksell’s analysis of the natural rate, we are compelled to reject these claims as false. As noted above, a definitive empirical test of Selgin’s central contention — that the super-low interest rates we are experiencing are not caused by expansionary monetary policy — would involve the termination of all open market operations. Somehow, I doubt Selgin would approve of such a test.